Methylglyoxal-Derived Hydroimidazolone Advanced Glycation End-Products of Human Lens Proteins

Abstract

purpose. To determine the concentrations of methylglyoxal-derived advanced glycation end-products (AGEs), the hydroimidazolones MG-H1 and -H2, in soluble human lens proteins and compare them with the concentrations of other methylglyoxal-derived AGEs and pentosidine.

methods. Lens protein samples were hydrolyzed enzymatically. AGEs were assayed without derivatization by HPLC with tandem mass spectrometry; the fluorescent AGEs argpyrimidine and pentosidine were assayed by fluorometric detection. MG-H1 and -H2 were resolved and assayed by fluorometric detection after derivatization with 6-aminoquinolyl-N-hydroxysuccimidylcarbamate (AQC).

results. The methylglyoxal-derived hydroimidazolones MG-H1 and -H2 were detected and quantified in human lens proteins. AGE concentrations (mean ± SEM) were: MG-H1 4609 ± 411 pmol/mg protein, MG-H2 3085 ± 328 pmol/mg protein, argpyrimidine 205 ± 19 pmol/mg protein, and pentosidine 0.693 ± 0.104 pmol/mg protein. The concentration of MG-H1 in human lens protein correlated positively with donor age (correlation coefficient = 0.28, P < 0.05), the concentration of MG-H2 (correlation coefficient = 0.78, P < 0.001) and argpyrimidine (correlation coefficient = 0.42, P < 0.01). The concentrations of AGEs were increased in cataractous lenses in comparison with noncataractous lenses: the increases were MG-H1 85%, MG-H2 122%, argpyrimidine 255%, and pentosidine 183% (P < 0.001). Multiple logistic regression analysis showed a significant link of cataract to donor age (regression coefficient β = 0.094, P = 0.026) and argpyrimidine (β = 0.022, P = 0.002).

conclusions. Methylglyoxal hydroimidazolones are quantitatively major AGEs of human lens proteins. These substantial modifications of lens proteins may stimulate further glycation, oxidation, and protein aggregation leading to the formation of cataract.

Methylglyoxal is a reactive α-oxalaldehyde metabolite and precursor of advanced glycation end-products (AGEs). In the lens, methylglyoxal is formed by the degradation of triosephosphates, ¹ the degradation of glycated proteins, ² and lipid peroxidation. ³ The concentration of methylglyoxal in the lens was relatively high (∼1–2 μM), ⁴ compared with the mean concentration of methylglyoxal in whole blood samples of normal healthy human subjects of approximately 80 nM. ⁵ Most methylglyoxal was detoxified by metabolism to d-lactate by the glyoxalase system. The glyoxalase system comprises two enzymes, glyoxalase I and glyoxalase II, and a catalytic amount of glutathione (GSH) cofactor. Glyoxalase I is a glutathione-dependent enzyme and catalyzes the conversion of methylglyoxal to S-d-lactoylglutathione. The glyoxalase system has been detected in mammalian lens. ^{6 7} In the human lens, the activity of glyoxalase I and the concentration of GSH decreased with donor age. ^{4 8} This suggests that the detoxification of methylglyoxal in the human lens in situ declines with age. Hence, the glycation of lens proteins and formation of AGEs from methylglyoxal may increase with age. 

Methylglyoxal reacted reversibly with cysteine residues to form hemithioacetal adducts, and lysine and arginine residues to form glycosylamine residues. ⁹ Further reaction with lysine residues occur irreversibly to form N _ε-(1-carboxyethyl)lysine (CEL) ¹⁰ and 1,3-di(N _ε-lysino)-4-methyl-imidazolium (MOLD). ¹¹ An irreversible reaction of methylglyoxal with arginine residues forms N _δ-(4-carboxy-4,6-dimethyl-5,6-di-hydroxy-1,4,5,6-tetra-hydropyrimidine-2-yl)ornithine (THP), ¹² and argpyrimidine, ¹³ but the major adduct in proteins modified minimally, as found in vivo, is methylglyoxal-derived hydroimidazolone (MG-H). MG-H is formed as three structural isomers: N _δ-(5-hydro-5-methyl-4-imidazolon-2-yl)-ornithine (MG-H1), 2-amino-5-(2-amino-5-hydro-5-methyl-4-imidazolon-1-yl)pentanoic acid (MG-H2), and 2-amino-5-(2-amino-4-hydro-4-methyl-5-imidazolon-1-yl)pentanoic acid (MG-H3) (Fig. 1) . MG-H3 is unstable and was not usually quantified. ^{14 15}  

In this report, we describe for the first time that methylglyoxal hydroimidazolones MG-H1 and -H2 are quantitatively major AGEs in the human lens. 

Materials and Methods

Human Lens Samples

Human cataractous lenses were obtained from patients in the Department of Ophthalmology, Friedrich-Schiller University (Jena, Germany) who were undergoing routine cataract surgery. Noncataractous lenses of patients with injury or trauma were used as the control. Further noncataractous lenses were obtained from donor eyes and assessed by one of the authors (GMH) through an arrangement with the Southern Eye Bank, Inc. (New Orleans, LA). Lenses were removed by an aseptic technique with a surgical loop within 8 hours of death. The lenses were examined in toto by transillumination and low-power microscopy. Color, clarity, and capsular integrity were determined. Medical history was noted. The lenses were stored at −84°C. Samples were dispatched on dry ice to the University of Essex for AGE analysis. Decapsulated individual lenses were homogenized on ice in 5 mL of phosphate-buffered saline (PBS) followed by centrifugation at 100,000g for 1 hour at 4°C. The supernatants containing the water-soluble lens proteins were stored at −80°C before analysis. 

Sixteen lenses without cataract were from donors of mean age 56 ± 10 years; 3 were women and 13 were men. Thirty-nine samples with cataract were from donors with a mean age of 71 ± 14 years, which was significantly higher than that of the noncataractous donors (P < 0.001); 20 were women and 19 were men. Twenty-six subjects had diabetes mellitus and 29 were nondiabetic, with mean ages of 66 ± 16 and 67 ± 12 years, respectively (P > 0.05). Cataract was graded by color of the lens protein (grades 1 to 3). Five subjects had grade 1 (light yellow lens, mean age, 64 ± 20 years), 37 had grade 2 (yellow-light brown, mean age, 65 ± 14 years), and 13 had grade 3 (dark yellow-brown, mean age, 74 ± 13 years), a significantly higher age than subjects with grade 2 color; P < 0.05). This research followed the tenets of the World Medical Association Declaration of Helsinki on ethical principles for medical research involving human subjects. 

Delipidification, Washing, and Enzymatic Hydrolysis of Lens Proteins

Lens protein samples (100–500 μg; 2 mg/mL) were extracted with a 2 × 1-volume of water-saturated diethyl ether to remove lipids, and the residual ether was removed by centrifugal evaporation. An aliquot of protein sample (∼500 μg) was then diluted to 500 μL with water and concentrated to approximately 50 μL by ultrafiltration (12-kDa cutoff membrane). This washing procedure was repeated a further two times. The protein concentration was then determined by the Bradford method. ¹⁶ Protein samples were hydrolyzed enzymatically under nitrogen, as described. ¹⁴ This hydrolysate was used in the assay of AGEs by the LC-MS/MS and fluorescence techniques. 

Assay of AGEs by the LC-MS/MS, Intrinsic Fluorescence, and AQC-Derivatization Chromatographic Methods

The concentration of MG-H, arginine, and lysine in soluble lens protein hydrolysates was determined by LC-MS/MS using the stable isotope-substituted standards for internal standardization with reference to calibration curves of authentic standards. [Guanidino-¹⁵ N ₂]-l-arginine, [¹³C₆]-l-lysine, (all 98% isotopic purity) were purchased from Cambridge Isotope Laboratories (Andover, MA). [Guanidino-¹⁵ N ₂]-MG-H1 was prepared from [guanidino-¹⁵ N ₂]-l-arginine after conversion to the N _α-t-BOC derivative. ^{14 17} Samples were assayed by LC-MS/MS using a separation module (model 2690) with a triple quadrupole mass spectrometric detector (Quattro Ultima; Waters-Micromass, Manchester, UK). Two 5 μm columns (Hypercarb; Thermo Hypersil, Ltd., Runcorn, UK) in series were used: 2.1 × 50 mm (column 1) and 2.1 × 250 mm (column 2). The mobile phase was 26 mM ammonium formate (pH 3.8), with a two-step gradient of acetonitrile (17–25 minutes, 0%–31% acetonitrile; 25–30 minutes, 31% acetonitrile). The flow rate was 0.2 mL/min. The flow was diverted to bypass column 2 at 20 minutes to facilitate elution of MG-H. Flow from the column during the interval of 4 to 30 minutes was directed to the MS/MS detector. Amino acids and AGEs were detected by electrospray-positive ionization-mass spectrometric multiple-reaction monitoring (MRM), with which the strongest fragment ion response, formed from a specific parent ion, was detected. The ionization source temperature was 120°C and the desolvation gas temperature, 350°C. The cone gas and desolvation gas flow rates were 150 and 550 L/h, respectively. The capillary voltage was 3.55 kV and the cone voltage, 80 V. Argon gas (2.7 × 10⁻³ mbar) was in the collision cell. Programed molecular ion and fragment ion masses and collision energies were optimized to ±0.1 Da and ±1 eV for MRM detection of analytes. Amounts of internal standard used were: 10 nmol for amino acids and 50 pmol for MG-H1. The retention times, MRM transitions (molecular ion > fragment ion masses), collision energy, and fragment losses for analytes and calibration standards were, respectively: lysine 5.0 minutes, 147.1 > 84.3 Da, 15 eV, H₂CO₂+NH₃, [¹³C₆]-lysine; arginine 10.9 minutes, 175.2 > 70.3 Da, H₂CO₂+NH₂C(⋕NH)NH₂, 15 eV, [¹⁵ N ₂]-arginine; MG-H 23.6 and 24.0 minutes (two epimers), 229.2 > 114.3 Da, 14 eV, NH₂CH(CO₂H)CH₂CH⋕CH₂ and [¹⁵ N ₂]-MG-H1. The concentration of the fluorescent AGEs argpyrimidine and pentosidine was determined by HPLC with fluorometric detection (model 2475 fluorometric detector; Waters). The column was a 5-μm particle size, with dimensions of 50 × 2.1 mm (Hypercarb; Waters). The mobile phase was 0.1% trifluoroacetic acid in 10% acetonitrile with a linear gradient to 50% acetonitrile at 15 minutes and isocratic 50% acetonitrile thereafter. The limits of detection for argpyrimidine and pentosidine were 400 fmol and 20 fmol, respectively. The recoveries of MG-H1, argpyrimidine, and pentosidine in enzymatic hydrolysis were 83%, 84%, and 101% and the recoveries of arginine and lysine compared with acid hydrolysis were 94%. 

MG-H structural isomers coeluted in LC-MS/MS analysis, and therefore the AQC derivatization technique that resolved MG-H isomers ^{14 15} enabled the analysis of MG-H1 and -H2. Aliquots of hydrolysate (50 μL, equivalent to 50 μg protein) were derivatized by AQC and analyzed by HPLC with fluorometric detection, as described. ^{14 15}  

The formation of AGEs in lens proteins by glycation with methylglyoxal in vitro was investigated by incubation of lens protein under conditions similar to those used to prepare human serum albumin modified minimally by methylglyoxal. ¹⁵ Lens protein (6 mg/mL) was incubated with and without methylglyoxal (500 μM) in sodium phosphate buffer (100 mM [pH 7.4] and at 37°C) for 24 hours. The samples were then washed by ultrafiltration and analyzed for AGEs, as described earlier. 

Results

Methylglyoxal-Derived Hydroimidazolones, Argpyrimidine, and Pentosidine in Human Lens Proteins

Methylglyoxal-derived AGEs in enzymatic hydrolysates of lens proteins were determined without derivatization by LC-MS/MS with stable isotope-substituted internal standardization. In the volatile buffer eluent method used, the epimer pairs (ld- and ll-stereoisomers) of the structural isomers MG-H1, -H2, and -H3 were partially resolved in the LC-MS/MS chromatogram, but the structural isomers coeluted. There were therefore two peaks reflecting the ld- and ll-epimers coeluting of all three structural isomers (Fig. 2a . The detector response to each isomer was not significantly different, and therefore MG-H determination by this method reflects the sum of all isomers. The internal standard [¹⁵ N ₂]MG-H1 coeluted with the MG-H isomers (Fig. 2b) . The MG-H/[¹⁵ N ₂]MG-H1 detector response ratio was calibrated with authentic MG-H1 (1–100 pmol). There were high levels of MG-H in human lens proteins (∼8 nmol/mg protein). 

To measure the concentrations of MG-H1 and -H2, we derivatized enzymatic hydrolysates of lens proteins by AQC and analyzed the resultant adducts by HPLC with fluorometric detection. ¹⁴ Each hydroimidazolone was detected as two peaks eluting with similar retention times (Fig. 3a) . These were epimers due to racemization of the chiral C₅ of the hydroimidazolone ring. MG-H1 gave two epimers eluting between proline and the internal standard (IS) α-aminobutyric acid in the HPLC chromatogram at retention times 81.0 and 83.4 minutes. MG-H2 yielded two epimers eluting immediately before and after the IS at retention times 85.8 and 91.7 minutes. The chromatographic peaks of the second epimer of MG-H1 and the first epimer of MG-H2 were well resolved from other peaks in the chromatographic analysis of lens protein hydrolysates, and hence these peaks were integrated for quantification of MG-H1 and -H2. There were high levels of MG-H1 and -H2 in human lens proteins. The chromatographic peaks of MG-H1 and -H2 AQC adducts were clearly discernible in the analytical chromatograms (Figs. 3b 3c) . MG-H1 was the major isomer. Estimation of MG-H by the LC-MS/MS method provided data that corroborated well with the sum of MG-H1 and -H2 estimated by the AQC assay, indicating that MG-H3 contributed little to the detection of MG-H. This was expected, because MG-H3 is a minor product of glycation under physiological conditions. ^{14 15} Lens proteins glycated in vitro with 500 μM methylglyoxal showed a marked increase in both epimers of MG-H1 and the presence of a further two peaks due to THP (Fig. 3d) . THP was below the limit of detection in lens proteins glycated in vivo. 

The concentrations of argpyrimidine and pentosidine in lens protein hydrolysates were determined with high sensitivity by HPLC with fluorometric detection. The mean concentrations of MG-H1 and -H2 in lens proteins were approximately 6700 and 4500 times higher than pentosidine and 22 and 15 times higher than argpyrimidine. The concentration of argpyrimidine was approximately 300 times higher than pentosidine (Table 1) . Expressed as percentage of arginine residues, AGE concentrations were: MG-H1 1.5%, MG-H2 1.0%, argpyrimidine 0.068%, and pentosidine 0.00023%. 

The concentration of MG-H1 in human lens protein correlated positively with donor age (correlation coefficient = 0.28, P < 0.05). The correlation was not significant for diabetic subjects. Linear regression of MG-H1 on donor age for nondiabetic subjects gave the regression equation: MG-H1 (pmol/mg protein) = 70 × donor age (years) + 705. The concentration of MG-H1 also correlated positively with the concentration of MG-H2 (correlation coefficient = 0.78, P < 0.001) and with the concentration of argpyrimidine (correlation coefficient = 0.42, P < 0.01; Fig. 4 ). There was no significant increase in AGE concentrations in lens proteins of diabetic subjects, relative to nondiabetic subjects. The concentrations of MG-H1, MG-H2, MG-H1+MG-H2, argpyrimidine, and pentosidine were increased in cataractous lenses, relative to noncataractous lenses (P < 0.001), but in cataractous lenses, the analyte concentrations were not increased significantly in lenses with intense versus mild coloration (Table 1) . Multiple logistic regression analysis to examine variables linked to cataract showed a significant link of cataract to donor age (regression coefficient β = 0.094, P = 0.026) and the concentration of argpyrimidine (β = 0.022, P = 0.002). 

Discussion

Methylglyoxal-derived hydroimidazolone MG-H was the major methylglyoxal-derived AGE in proteins modified minimally by methylglyoxal in vitro. Minor amounts of THP and argpyrimidine were also formed. ^{14 15} We therefore expected that MG-H would be the major methylglyoxal-derived AGE in lens proteins in vivo. In this report, we found that MG-H was indeed the major methylglyoxal-derived AGE and that it was present at high concentration. The estimates were deduced by corroborative LC-MS/MS and AQC derivatization techniques. From the results of the AQC assay of AGEs, it was clear that both structural isomers MG-H1 and -H2 were present. Enzymatic hydrolysis of proteins was essential, because acid hydrolysis led to the degradation of hydroimidazolone AGEs by approximately 90%. A further isomer of methylglyoxal-derived hydroimidazolone MG-H3 was also formed, but it degraded significantly during analysis and was not quantified. ^{14 15} Little MG-H2 is formed in proteins glycated in vitro during short incubation periods (24 hours). MG-H1 is the major adduct under these conditions of kinetic control of hydroimidazolone isomer formation. Indeed, when we glycated lens protein with 500 μM methylglyoxal in vitro (Fig. 3d) , MG-H1 and a small amount of THP were the major products formed. With glycation products forming over much longer periods in vivo, the MG-H isomers may attain thermodynamic equilibrium when MG-H2 is also present. Consistent with this, there was a highly significant correlation of MG-H1 with MG-H2 in lens proteins (Fig. 4b) . MG-H content of human lens proteins is much higher than other AGEs reported previously ¹⁸ and may be of functional and pathologic significance. 

The concentrations of MG-H1 and -H2 in lens proteins are similar in the extent of amino acid modification (1%–2% arginine) to that of fructosamine (1%–2% lysine). ¹⁸ This suggests that MG-H1 and -H2 are major glycation adducts quantitatively. Recently, we have found similar high concentrations of MG-H1 in protein extracts of human blood cells and tissues of laboratory rats. ¹⁹ Argpyrimidine and pentosidine were present at much lower concentrations (∼0.1% arginine and ∼0.0002% arginine in cataractous lenses, respectively). Other methylglyoxal-derived AGEs are also found in lens proteins at low concentrations: CEL 0.1%–0.4% lysine and MOLD 0.1%–0.8% lysine. ^{10 11 20} Hence, MG-H is the major methylglyoxal-derived AGE in the lens. The concentration of MG-H1 increased with donor age in the current study, as did the concentrations of CEL and MOLD reported previously. ¹⁰ The predicted increases in MG-H1, CEL, and MOLD over the subject age interval of 20 to 75 years are 183%, 92%, and 275%, respectively. 

MG-H1, MG-H2, and argpyrimidine are relatively short-lived AGEs with half-lives of only 1 to 2 weeks under physiological conditions. ¹⁴ The levels of these analytes in long-lived proteins therefore may reflect the fluctuation of methylglyoxal concentration in the 2 to 4 weeks before lens extraction. The concentrations of these AGEs in long-lived lens proteins reflect the steady state concentration of AGE achieved as a balance of the rates of AGE formation and degradation. When significant correlations of these AGE analytes are found with donor age, factors influencing the steady state AGE concentration that change slowly over many years are likely to produce the correlation. These factors are GSH concentration, glyceraldehyde-3-phosphate dehydrogenase activity, and glyoxalase I activity. ^{4 8 21} The activity of glyoxalase I and the concentration of triosephosphates are important variables controlling methylglyoxal concentration and related glycation in cultured rat lens. ²²  

The association of protein glycation with cataract was further supported in this work by the finding of increased concentrations of MG-H1, argpyrimidine, and pentosidine in soluble protein extracts of cataractous lenses in comparison with noncataractous lens proteins. We found concentrations of argpyrimidine and pentosidine in lens proteins similar to those reported previously. ^{23 24} The increased glycation of proteins in cataractous lenses—particularly the higher extent of glycation by MG-H—may induce protein conformational changes that stimulate further glycation and oxidation and trigger protein aggregation leading to cataract. To study the association of variables with cataract formation, a multiple logistic regression model was computed. This indicated that cataract was associated significantly with donor age and argpyrimidine concentration. The formation of argpyrimidine is favored by high concentrations of methylglyoxal ¹⁵ and oxidative processes. ¹³ Argpyrimidine may also be degraded by oxidative processes. ¹⁴ The link of argpyrimidine to cataract in the multiple logistic regression model suggests that the formation of argpyrimidine is either involved in the development of cataract or is a surrogate indicator of other critical factors in cataractogenesis. The lack of significant increases in AGE concentrations in diabetic lenses may have been due to masking of diabetes-associated changes by the effects of donor age and oxidative stress. 

We conclude that MG-H is a major AGE in human lens proteins quantitatively. The modification of lens crystallins by methylglyoxal led to a decrease in arginine residues and loss of positive charge. ²⁵ The formation of MG-H1 and -H2 are likely adducts producing this effect with up to 2% of total arginine modified. There are 10 to 20 arginine residues in the human crystallin isoforms. There is therefore an expectation that 10% to 20% of crystallin molecules have a MG-H modification. Loss of a single arginine residue may promote cataract formation, as found in the R58H mutation in γ-d-crystallin. ^{26 27} Changes in lens protein charge due to MG-H formation, although short lived, may induce protein refolding and stimulate long-lived irreversible modifications, such as oxidation and proteolysis, that are major protein modifications implicated in cataractogenesis. ²⁸ Further studies are needed to investigate the involvement of hydroimidazolone AGEs in cataract formation. 

 Supported by a contract from the Institute for Pharmaceutical Discovery (IPD), Branford, Connecticut; and the Wellcome Trust, United Kingdom (PJT, NA, GMH).

 Submitted for publication June 7, 2003; revised July 14, 2003; accepted July 24, 2003.

 Disclosure: N. Ahmed, None; P.J. Thornalley, None; J. Dawczynski, None; S. Franke, None; J. Strobel, None; G. Stein, None; G.M. Haik, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Methylglyoxal-derived AGEs.

Figure 2.

 

Detection of methylglyoxal-derived hydroimidazolone in human lens protein by HPLC with tandem mass spectrometry. (a) MG-H; (b) [¹⁵ N ₂]MG-H1 (50 pmol) in human lens protein (50 μg equivalent).

Figure 3.

 

Analysis of methylglyoxal-derived hydroimidazolones in human lens proteins by the chromatographic AQC method. (a) Methylglyoxal-derived hydroimidazolones (500 pmol MG-H1, MG-H2, and the IS α-aminobutyric acid) and amino acid standards (10 nmol) by the chromatographic AQC method. (b) Noncataractous lens; (c) cataractous lens; and (d) the noncataractous lens protein in (a) reanalyzed after incubation with 500 μM methylglyoxal for 24 hours at pH 7.4 and 37°C in vitro.

Table 1.

 

Concentrations of the Methylglyoxal-Derived Hydroimidazolones Argpyrimidine and Pentosidine in Proteins of Human Lens

A. Comparison of Clear and Cataractous Lenses
	Analyte	AGE (pmol/mg protein)
			Clear (n = 16)			Cataractous (n = 39)
			Mean ± SEM	Range	Mean ± SEM	Range
MG-H1	2,848 ± 328	1253–6,361	5,278 ± 496	400–14,380
MG-H2	1,504 ± 204	775–3,511	3,348 ± 288	161–7940
MG-H1+MG-H2	4,352 ± 857	2,029–9,872	8,626 ± 733	1220–22,320
Argpyrimidine	72 ± 21	2–285	256 ± 20	84–567
Pentosidine	0.30 ± 0.04	0.08–0.64	0.85 ± 0.13	0.12–3.80

B. Comparison of Cataractous Lenses with Mild and Intense coloration
	Analyte	AGE (pmol/mg protein)
			Cataractous, Mild Coloration (n = 24)			Cataractous, Intense Coloration (n = 13)
			Mean ± SEM	Range	Mean ± SEM	Range
MG-H1	5,878 ± 743	400–14,380	4,049 ± 434	2,573–7,945
MG-H2	3,702 ± 386	820–7,940	2,710 ± 455	161–5,099
MG-H1+MG-H2	9,580 ± 1,076	1,220–22,320	6,759 ± 745	2,985–13,044
Argpyrimidine	278 ± 27	88–567	221 ± 29	84–478
Pentosidine	0.94 ± 0.20	0.12–3.80	0.75 ± 0.14	0.14–2.13

 Significance: clear versus cataractous, for all analytes, P < 0.001; cataractous, mild versus intense coloration, for all analytes, P > 0.05. An outlier in the noncataractous group (MG-H1 = 13,660 pmol/mg protein and MG-H2 = 14,840 pmol/mg protein) was omitted from the statistical analysis. Two cataractous lenses had no discernible coloration.

Figure 4.

 

Correlation analysis of methylglyoxal-derived AGEs in human lens proteins. (a) MG-H1 and donor age (correlation coefficient = 0.28, P < 0.05; nondiabetic subjects only), with the regression equation MG-H1 (pmol/mg protein) = 70 × donor age (years) + 705; (b) MG-H2 and MG-H1 (correlation coefficient = 0.78, P < 0.001), with the regression equation MG-H2 (pmol/mg protein) = 0.47 × MG-H1 (pmol/mg protein) + 674; and (c) argpyrimidine and MG-H1 (correlation coefficient = 0.42, P < 0.01) with the regression equation argpyrimidine (pmol/mg protein) = 0.021 × MG-H1 (pmol/mg protein) + 105.

The authors thank the Southern Eye Bank Inc., New Orleans, Louisiana, for assistance with the research and Sinan Battah for assistance in the preparation of [¹⁵ N ₂]MG-H1.